Skip to main content

Browse the archive

Revision e009d5dc0a94a7133e5f1c083732d760bfd038e6 authored by Michal Hocko on 12 March 2015, 23:25:52 UTC, committed by Linus Torvalds on 13 March 2015, 01:46:07 UTC 
mm, oom: do not fail __GFP_NOFAIL allocation if oom killer is disabled
 
Tetsuo Handa has pointed out that __GFP_NOFAIL allocations might fail
after OOM killer is disabled if the allocation is performed by a kernel
thread.  This behavior was introduced from the very beginning by
7f33d49a2ed5 ("mm, PM/Freezer: Disable OOM killer when tasks are frozen").
 This means that the basic contract for the allocation request is broken
and the context requesting such an allocation might blow up unexpectedly.

There are basically two ways forward.

1) move oom_killer_disable after kernel threads are frozen.  This has a
   risk that the OOM victim wouldn't be able to finish because it would
   depend on an already frozen kernel thread.  This would be really tricky
   to debug.

2) do not fail GFP_NOFAIL allocation no matter what and risk a
   potential Freezable kernel threads will loop and fail the suspend.
   Incidental allocations after kernel threads are frozen will at least
   dump a warning - if we are lucky and the serial console is still active
   of course...

This patch implements the later option because it is safer.  We would see
warning rather than allocation failures for the kernel threads which would
blow up otherwise and have a higher chances to identify __GFP_NOFAIL users
from deeper pm code.

Signed-off-by: Michal Hocko <mhocko@suse.cz>
Acked-by: David Rientjes <rientjes@gooogle.com>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Cc: Tetsuo Handa <penguin-kernel@i-love.sakura.ne.jp>
Cc: "Rafael J. Wysocki" <rjw@rjwysocki.net>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
1 parent 8792f77
Raw File
ulist.h 
/*
 * Copyright (C) 2011 STRATO AG
 * written by Arne Jansen <sensille@gmx.net>
 * Distributed under the GNU GPL license version 2.
 *
 */

#ifndef __ULIST__
#define __ULIST__

#include <linux/list.h>
#include <linux/rbtree.h>

/*
 * ulist is a generic data structure to hold a collection of unique u64
 * values. The only operations it supports is adding to the list and
 * enumerating it.
 * It is possible to store an auxiliary value along with the key.
 *
 */
struct ulist_iterator {
#ifdef CONFIG_BTRFS_DEBUG
	int i;
#endif
	struct list_head *cur_list;  /* hint to start search */
};

/*
 * element of the list
 */
struct ulist_node {
	u64 val;		/* value to store */
	u64 aux;		/* auxiliary value saved along with the val */

#ifdef CONFIG_BTRFS_DEBUG
	int seqnum;		/* sequence number this node is added */
#endif

	struct list_head list;  /* used to link node */
	struct rb_node rb_node;	/* used to speed up search */
};

struct ulist {
	/*
	 * number of elements stored in list
	 */
	unsigned long nnodes;

	struct list_head nodes;
	struct rb_root root;
};

void ulist_init(struct ulist *ulist);
void ulist_reinit(struct ulist *ulist);
struct ulist *ulist_alloc(gfp_t gfp_mask);
void ulist_free(struct ulist *ulist);
int ulist_add(struct ulist *ulist, u64 val, u64 aux, gfp_t gfp_mask);
int ulist_add_merge(struct ulist *ulist, u64 val, u64 aux,
		    u64 *old_aux, gfp_t gfp_mask);

/* just like ulist_add_merge() but take a pointer for the aux data */
static inline int ulist_add_merge_ptr(struct ulist *ulist, u64 val, void *aux,
				      void **old_aux, gfp_t gfp_mask)
{
#if BITS_PER_LONG == 32
	u64 old64 = (uintptr_t)*old_aux;
	int ret = ulist_add_merge(ulist, val, (uintptr_t)aux, &old64, gfp_mask);
	*old_aux = (void *)((uintptr_t)old64);
	return ret;
#else
	return ulist_add_merge(ulist, val, (u64)aux, (u64 *)old_aux, gfp_mask);
#endif
}

struct ulist_node *ulist_next(struct ulist *ulist,
			      struct ulist_iterator *uiter);

#define ULIST_ITER_INIT(uiter) ((uiter)->cur_list = NULL)

#endif
back to top